Results
metrics
recommend length:15- accuracy: 0.8630
- precision:0.8661
- recall:0.8630
- f1: 0.8619
- roc_area:0.8564
- confusion matrix:
-
Predicted 0 1 Actual 0 37 3 1 7 26
1. Forward feature search (RandomForest cv=5）
2. Comparison dimensionality reduction(left:raw dataset ,right:use MRMD3.0)
3. top 15 features distribution
4. top 15 features Pearson correlation coefficient
5. top 15 feature score PageRank
feature ranking
| NO. | features name | score |
|---|---|---|
| 1 | Y | 0.05879620905677517 |
| 2 | K | 0.05597976215208717 |
| 3 | E | 0.05425100824116193 |
| 4 | N | 0.05421276482680626 |
| 5 | V | 0.0537541248793991 |
| 6 | S | 0.05333776398924654 |
| 7 | I | 0.05317466321392428 |
| 8 | P | 0.05263539302067472 |
| 9 | Q | 0.05152564063831892 |
| 10 | G | 0.0505196168186521 |
| 11 | D | 0.049194750159629284 |
| 12 | R | 0.0487031018830841 |
| 13 | C | 0.04854651300809624 |
| 14 | F | 0.047853961195145574 |
| 15 | A | 0.047722207896049404 |
| 16 | H | 0.047373544807334585 |
| 17 | M | 0.04705103848134251 |
| 18 | W | 0.04437020012917783 |
| 19 | L | 0.04377576122280248 |
| 20 | T | 0.03722197438029176 |